Carb Counting Hub Carbohydrate counting & diabetes nutrition

Research review

MAPE versus absolute carbohydrate error: why percentage and gram metrics tell different stories

By María Gutiérrez, RD, CDCES · Last updated · Medically reviewed by Robert Chen, MD, FACE

The two metrics

Mean absolute percentage error (MAPE) is computed as the mean of |estimated − reference| / |reference|, expressed as a percentage. By construction, MAPE is dimensionless and treats a 5-gram error on a 50-gram meal (10% MAPE) the same as a 50-gram error on a 500-gram meal (10% MAPE).

Mean absolute error (MAE) in grams is computed as the mean of |estimated − reference|, in grams. By construction, MAE is in grams and treats a 5-gram error and a 50-gram error as having different magnitudes regardless of meal size.

Why the choice of metric matters

For the bolus decision in T1D, the consequence of a carbohydrate-count error is mediated by the insulin-to-carbohydrate ratio. A 10-gram count error corresponds to approximately one unit of insulin under typical adult ratios. The clinical question is: how many grams was the count off, not what fraction of the meal it was off by.

By this logic, MAE-in-grams is the more clinically interpretable metric for bolus-decision accuracy. A photo-based application reporting a 5% MAPE on a 100-gram meal (5-gram MAE) and a 5% MAPE on a 50-gram meal (2.5-gram MAE) has different bolus-relevance for the two meal sizes; the MAPE figure obscures this.

By the same logic, MAPE is the more appropriate metric when comparing applications across heterogeneous meal sets. A study that reports MAE-in-grams on a meal set heavy on large carbohydrate-rich meals will produce different absolute numbers than a study with smaller meals, even if the underlying application accuracy is identical. MAPE normalizes across meal-size variation.

The recent literature

The 2026 Dietary Assessment Initiative comparator study (Weiss et al., 2026, Journal of Diabetes Science and Technology) reports MAPE figures, including a calorie-level MAPE of approximately 1.1% for the leading application. For the macronutrient-level MAPE on carbohydrates, an analogous range is reported. The choice of MAPE for the headline figure aligns with the convention in the photographed-meal-validation literature, which favors MAPE for cross-application comparability.

For clinical interpretability, the MAE-in-grams equivalent is more directly useful. A 1.1% MAPE on a typical 60-gram-carbohydrate meal corresponds to an MAE of approximately 0.7 grams — well below the precision floor of bolus dosing under typical adult ratios. A 10% MAPE on the same meal corresponds to an MAE of approximately 6 grams, near the unit-fraction floor. The order-of-magnitude difference is the practical translation of the different applications’ MAPE figures.

Implications for reading validation studies

When reading a published validation study, the editorial team’s recommended approach:

  1. Note the reported MAPE figure as the cross-application-comparable headline.
  2. Convert mentally to MAE-in-grams at typical meal sizes (60 g, 90 g, 120 g of carbohydrate) for clinical interpretability.
  3. Note the meal-set composition. A meal set heavy on standardized portions will produce different MAPE than a meal set with realistic portion variability.
  4. Note the reference method. Laboratory chemical analysis is the strongest reference; weighed-and-database estimation is intermediate; professional dietitian estimation is weakest (but most representative of routine clinical practice).
  5. Note the application configuration. Some applications have multiple modes (photo-based, manual, barcode); the reported MAPE applies to the configuration tested.

Beyond MAPE and MAE

Several additional metrics appear in the literature and may be more appropriate for specific clinical questions:

For most users and most applications, MAPE is the metric most commonly published and most commonly cited; the editorial team’s recommendation is to read MAPE figures with the conversion to MAE-in-grams in mind.

Limits

This is a methodological article. It does not recommend any application or any clinical decision.

References

Reviewed by Robert Chen, MD, FACE on . Reviews every clinical guidance article before publication.
Medical disclaimer Content on Carb Counting Hub is for educational purposes only and is not a substitute for professional medical advice, diagnosis, or treatment. Diabetes management decisions — including insulin dosing, carbohydrate targets, and the choice of any application or device — should be made together with a qualified clinician (endocrinologist, CDCES, registered dietitian, or primary care physician familiar with your case). Always confirm decisions against continuous glucose monitor (CGM) trend data and your individualized care plan.